BAFFI Center on International Markets, Money and Regulation

Transcription

1 BAFFI Center on International Markets, Money and Regulation BAFFI Center Research Paper Series No CENTRAL BANKING, MACROPRUDENTIAL SUPERVISION AND INSURANCE By Donato Masciandaro and Alessio Volpicella This Paper can be downloaded without charge from The Social Science Research Network Electronic Paper Collection: Electronic copy available at:

2 CENTRAL BANKING, MACROPRUDENTIAL SUPERVISION AND INSURANCE Donato Masciandaro and Alessio Volpicella Abstract This paper offers a positive analysis of the relationships between central banking, macro prudential supervision and insurance, presenting three contributions. Starting from the review of the recent theoretical models, which directly or indirectly take the issue of the central bank involvement in macro supervision, it has been discovered that two relevant topics are missed: the political economy of the macro supervisory design and the role of the insurance interconnectedness. Therefore a principal agent model is used to design a political economy framework, which explains how the politicians can shape the central bank governance in addressing systemic risks in insurance. Finally, the framework is used to analyse the present institutional settings in 39 countries. Keywords: Central Banking, Macro Prudential Supervision, Insurance Interconnectedness, Political Economy, Principal Agent. JEL Classification: E31, E52, E58 Department of Economics and Paolo Baffi Centre, Bocconi University and SUERF. Paolo Baffi Centre, Bocconi University. Electronic copy available at:

3 I. INTRODUCTION After the 2008 financial turmoil considerable efforts has been made to draw lessons for the design of the supervisory architectures. One of the main issues, which have been and still are addressed, is how to prevent the instability of the financial system as a whole, building up a macro prudential framework. By definition, the macro prudential framework has to address the cross sectional dimension which characterized any systemic risk distribution. Therefore, two key features of any macro prudential architecture are: on the one side, how to define its governance - i.e. which authorities have to been involved in order to assign the set of powers and tools that characterized the macro prudential policy and on the other hand, how to indentify its perimeter - i.e. which are boundaries and features of the financial area that has to be supervised. The governance setting is a cornerstone in the building up of the macro prudential architecture and a crucial challenge in this area is to define the most effective involvement of the central bank, which tends to have a more or less leading role 1. Which are the drivers that can explain the position of the central bank into the macro supervisory framework? At the same time, the identification of the leading authorities implies a parallel challenge in understanding how their macro supervisory powers can be applied in the different sectors of the financial industry 2, including insurance. The macro prudential supervisor tends to include insurance into the perimeter of his/her responsibilities. How to explain the deepening of the macro supervisory power in the insurance sector? In this paper we intertwine the two research questions, offering a positive analysis of the relationships between central banking, macro prudential supervision and insurance. We present three contributions. Firstly, the complete review of the recent theoretical models which directly or indirectly take the issue of the central bank involvement in macro supervision discovers that two relevant topics are missed: the political economy of the macro supervisory design and the role of the insurance interconnectedness. Therefore a principal agent model is used to design a political economy framework, which explains how the politicians can shape the central bank governance in addressing systemic risks in insurance. Finally, the framework is used to analyse the present institutional settings in 39 countries. II. CENTRAL BANKING, MACRO PRUDENTIAL SUPERVISION AND INSURANCE: THE STATE OF THE ART In this paragraph, we focus on the review of central banking, macro prudential supervision and insurance in the recent economic literature. In order to shed light on these issues, the main models made after the Great Crisis are presented and discussed. Each of the following subparagraphs 1 International Monetary Fund (2011). 2 Lim et al. (2011)., IMF (2011) and (2013). 2 Electronic copy available at:

4 deals with a model, while the last one concludes and sums up the most important lessons we learn from the current literature. For each model, we analyze the aim, the theoretical framework, the empirical structure and its results (if any), the policy implications, the role of the central bank, financial and insurance markets (if any). II.I COUNTERCYCLICAL MACRO PRUDENTIAL POLICIES IN A SUPPORTING ROLE TO MONETARY POLICY Aim: N Diaye (2009) studies how macro prudential policies are able to sustain the monetary policy in keeping the output variations under control and maintaining the financial stability. The framework relies on a traditional model of monetary policy with a financial sector. The main outcome is that countercyclical macro prudential policies can be helpful in reducing output movements and drive down the risk of the financial system. In particular, countercyclical prudential measures allow the monetary authority to pursue output and inflation goals with smaller fluctuations in interest rates. Also, countercyclical tools are able to decrease volatility in asset prices and lean against a financial accelerator process, meaning that risks of the system decrease. In a fixed exchange rate regime countercyclical monetary policy is not possible macro prudential policies help to reduce a run-up asset prices and support the output stability. Theoretical framework: The framework merges a model including forward-looking indicators of balance sheet risk with a standard multi-country monetary model. In order to capture the role of credit and balance sheets conditions in the channelization of shocks financial accelerator process, borrowing costs rely on an external financial premium based on the Contingent Claims Analysis (CCA). The latter means that: (i) the risk is modelled as interactions between balance sheets; (ii) firms and households financial distress impacts on the banks balance sheets that, in turn, affect real economy. The CCA assumption merges balance sheets information with finance and risk management tools to build market-to-market balance sheets representing the risk better. The main point is that economic players firms, households and banks liabilities depend on asset values used as collateral. Such values are stochastic. The consequent volatility leads to uncertainty and, as a result, financial distress. In particular, the latter, i.e. the default, happens if the asset values fall below a certain promised payments the distress barrier, that is modelled as short term debt, a share of long term debt and one year of interest payments. It is noteworthy that creditors represent a de facto guarantee, because they are forced to absorb the losses in case of default. The distress risk is computed by using option pricing tools, while the guarantee is a put option on the assets used as collateral, where the exercise price is equal to the value of debt. The model includes an IS curve, a Philips curve, a version of Okun s law, a monetary policy rule, a yield curve, a modified uncovered interest parity, a labour income relationship. The supply side variables just follow a stochastic process. The financial side is based on the CCA assumption 3

5 tailored to firms, households and banks. Note that the model can allow up to four economies. The monetary policy has the goal to pursue the price stability and reduce the output variations; supervisory authority assures financial stability. The monetary tool is the interest rate that impacts on the exchange rate through a modified version of the uncovered interest parity and a short term market rate. The IS curve allows the exchange rate to move the aggregate domestic and foreign demand, while the Philipp curve allows the exchange rate to affect the inflation. Likewise, the short term market rate impacts on long term market rate through the yield curve that, in turn, has influence on the aggregate demand. Fluctuations in the aggregate demand move the output gap that, in turn, impacts on actual and expected inflation. Variations in monetary policy rate, market interest rates, exchange rates, output, unemployment and inflation impact on balance sheets of local and foreign firms, households and banks. Balance sheets changes affect the spreads difference between policy and market rates, that is the cost at which firms, households and bank can borrow. Also, the macroeconomic variables above impact on asset prices. In the model, supervisory measures are countercyclical and have the objective to limit the leverage of firms, households and banks. In particular, the latter are constrained by capital adequacy ratio equity over assets, while firms and households face loan to value ratios. Finally, the model is adaptable to represent a fixed exchange rate regime, wherein monetary policy is constrained. Empirical structure: The model simulates an exogenous increase in domestic demand and analyses how output and asset prices react with and without countercyclical macro prudential policies. Empirical results: First, consider the case where countercyclical measures are not implemented. Following the increase in domestic demand, the banks expand lending and support the increase in asset prices and drive up the collateral of firms and households. Second, consider prudential rules. Following the new capital requirements, the banks need to water down bank equity assets less risky debt leading to higher borrowing costs for the banks. As a result, financial conditions are tightened and output growth decelerates. Moreover, as output slows and spread increases, the asset prices boom is mitigated relative to the non-prudential rules world. Note that asset prices can also decrease depending on capital requirements calibration. Also, the required raise in monetary policy rate to target the variance in output and inflation is reduced by increase in spreads. In a fixed exchange rate economy, tightening in prudential rules leads to a lower-inflation scenario both in goods and assets markets and lessen output variations. Broadly speaking, empirical outcomes show that countercyclical macro prudential regulation tends to mitigate asset prices boom and stabilise output fluctuations. 4

6 Policy implications: The main policy implication is that countercyclical macro prudential measures are able to reduce output movements and decrease the risk of the financial stability. In particular, countercyclical prudential measures allow the monetary policy to reach output and inflation purposes with smaller variations in interest rates. Also, countercyclical tools are able to decrease volatility in asset prices and lean against a financial accelerator process, leading to a risk reduction. In a fixed exchange rate economy, countercyclical macro prudential policies reduce a runup asset prices and support the output stability. Role of the Central Bank: In the model, the central bank is the monetary authority, but prudential measures are implemented by another agency. Nevertheless, nor the possible role of the central bank in macro prudential supervision neither cooperation issues are debated. It is noteworthy that countercyclical prudential policies can help the central bank to stabilise output and inflation with smaller variation in interest rates relative to only-monetary policy world. In particular, such effects are evident in an economy wherein central bank is not able to implement restrictive policies fixed exchange rate regime. Financial and Insurance markets: The paper considers the financial system as a whole and does not segment it. II.II MONETARY AND MACRO PRUDENTIAL POLICY RULES IN A MODEL WITH HOUSE PRICE BOOMS Aim: Kannan, Rabanal, and Scott (2009) show that monetary policy rules that take into account credit and house prices growth can stabilise the economy better than what classical Taylor rule does; they use a DSGE model including housing and banking sector. Also, if macro prudential tools are considered to mitigate financial shocks, stabilisation gains clearly emerge. Nevertheless, under TFP shocks the optimal policy is not to use the macro prudential tool. As a result, a key factor in implementing the optimal monetary and macro prudential measures is understanding the source of the shocks. Theoretical framework: The framework is a DSGE model with some modifications relative to standard New Keynesian models: the model includes housing and financial markets. There are three main traits of the model. Firstly, households decide how to allocate the wealth in housing and nondurable goods. In particular, housing provides utility and allows wealth accumulation. Secondly, as the model assumes that direct financing is not allowed, financial intermediaries collect money from savers and lend it to borrowers with a spread depending on debtors net worth. Also, note that the profits from intermediation are transferred to the depositors. Thirdly, lending rate is a spread over the policy rate and depends on different factors: loan to value ratios, banks mark up, use of macro prudential tool by policymakers. Typical New Keynesians features of the model include: two sectors 5

7 model (durables and nondurables goods) under monopolistic competition and nominal rigidities; prices are sticky; habit preferences and adjustments costs are implemented to slow down consumption and investments, respectively; given a worker, shifting from production to building houses is costly; durable and nondurable goods production does not require capital and the economy is closed. Four monetary policy regimes are considered: (i) a classical Taylor rule, wherein central bank minimises both CPI inflation and output gap variance; (ii) an augmented Taylor rule, in which the central bank takes into account the growth rate of nominal credit, aside from output gap and CPI inflation; (iii) an augmented Taylor rule with a macro prudential tool. In particular, macro prudential tool directly affects the lagged nominal credit changes, meaning that macro prudential policymakers can impact on spreads fluctuations; regime (iv) is actually a modified version of (iii), wherein the coefficients of the Taylor rule, including the macro prudential tool, are optimized and function as a robustness check. In order to rank the monetary regimes, the paper considers a welfare criterion in which the policymaker cares about minimising inflation and output gap. Empirical structure: Calibration is based on matching the standard deviation of the following time series for the US: consumption, residential investment, CPI inflation, nominal house price inflation, short-term deposit rates, spreads between deposit and lending rates, nominal credit growth. All data are quarterly: consumption and residential investments come from the Bureau of Economic Analysis, CPI inflation from the Bureau of Labour Statistics and nominal house prices from the OECD. The short-term deposit rate is the 3-month T-bill rate, spreads are equal to the difference between the rate on all mortgage loans closed and the 3-month T-bill rate (both from the Haver database). Nominal credit growth relies on household credit market debt data obtained from the Flow of Funds of the Federal Reserve Board. In calibrating real and nominal rigidities, authors follow Iacoviello and Neri (2010); nevertheless, adjustments are implemented to match data second moments. Finally, in order to test the monetary regimes, three different shocks are simulated: financial shock (relaxation in lending standards), housing demand shock and TFP shock. Empirical results: First, consider a financial shock. In regime (i), the monetary policy simply reacts to increase in output gap and inflation by raising the policy rates. Under scenario (ii), the central bank reacts to credit growth, aside from output gap and inflation. In terms of volatility of macroeconomic variables, case (ii) performs better than (i). Nevertheless, under regime (iii), the outcomes are even better. In that case, since the macro prudential tool allows to act directly against the loosening in the borrowing conditions, then the monetary policy does not need to react strongly, meaning that better stability and less volatility in macroeconomic variables is reached. 6

8 Second, suppose that a housing demand shock occurs. Like for the financial shock, regimes (ii) and (iii) turn out to be better than case (i) in terms of stability. By comparing scenarios (ii) and (iii), note that output gap is more stable in case (ii), but inflation volatility is less in regime (iii). As a result, under housing demands shocks, ranking (ii) and (iii) is ambiguous and depends on welfare criteria of the policymaker and the relative weight given to the variance of output gap and inflation. Third, consider a TFP shock. Among the first three regimes, case (ii) performs better. As a matter of fact, using macro prudential rule regime (iii) increases downward pressure on CPI and output gap, leading to a less expansionary monetary policy than cases (i) and (ii) do. As a result, output gap and inflation volatility are greater in regime (iii). When all shocks are considered, regime (ii) is better; it is not surprising, given that TFP is the most common driver of business fluctuations, unless financial or housing demand shocks turn out to be dominant. Changing policymakers preferences in the welfare criterion and implementing robustness checks in calibrating the model confirms the results above. In particular, the optimal coefficients of Taylor rule and macro prudential tool regime (iv) support the main results of the model. Policy implications: Given the results of the model, some policy implications arise. First, including credit growth considerations into the monetary policy can help facing financial accelerator. Second, specific macro prudential tools can be useful to smooth credit cycles. In particular, macro prudential measures turn out to be welfare-improving under credit and housing shocks, but are not efficient under TFP shocks. As a result, some policy mistakes can occur: when the source of the shock is the productivity, constraining credit growth using macro prudential tools would drive down welfare, meaning that rigid and inflexible policy responses could lead to suboptimal results. That is particularly true by taking into consideration that TFP shocks usually cover most part of business cycles fluctuations and how is difficult to understand the actual source of the shocks. Only if policymaker was able to distinguish the shocks source, then a macro prudential tool would be useful, otherwise a monetary police regime including credit growth would be the preferred option. By summing up, the policymaker should be able to manage different reaction functions to face different shocks, instead of acting inflexibly; nevertheless, for the reasons above, such a task is not easy. Role of the Central Bank: In the model, the central bank is the monetary authority and, if macro prudential measures are implemented, is also in charge of macro supervision. Nevertheless, the paper does not deal with the consequences of unifying the monetary and macro prudential authorities and does not face cooperation issues. As the model shows, central bank should use macro prudential tool only when a financial shock occurs, while under TFP disturbances monetary policy with credit growth considerations is welfare-improving. As shown above, central bankers capacity to understand the source of the shocks plays a key role. 7

9 Financial and Insurance markets: Financial sector is not segmented. Authors model financial sector as intermediaries that collect money from savers and lend it to borrowers with a spread depending on debtors net worth. II.III LEANING AGAINST BOOM-BUST CYCLES IN CREDIT AND HOUSING PRICES Aim: Lambertini, Mendicino and Punzi (2011) aim at studying the gains in implementing monetary and macro supervision policies that lean against news-driven boom-busts cycles in credit and housing prices created by expectations of macroeconomic developments. The first outcome is that there is no trade off between the purposes of monetary policy and leaning against boom-bust cycles. Following a policy rate rule that strictly stabilises inflation is not efficient. In opposition to this, implementing a rule wherein interest rates react to financial variables helps smoothing cycles and increases welfare. Secondly, counter cyclical loan to value (LTV) ratios do not drive up the inflation and are more efficient in keeping the financial system stable than what monetary policy reacting to financial conditions does. Nevertheless, creditors and debtors heterogeneity in welfare criterion does not allow to rank the two regimes consistently. Theoretical framework: The framework is a DSGE model following Iacoviello and Neri (2010) and Lambertini, Mendicino and Punzi (2010). The agents operating in the economy are households, firms producing nondurable goods and housing, retailers and a central bank. The households can be savers or borrowers and both of them work in the production of goods and housing, consume and accumulate housing. In order to produce nondurable goods and housing, firms use capital, land, intermediate inputs and labour supplied by households. As a result, firms provide households and savers with wages and repayment of rented capital, respectively. Savers own retailers who differentiate final goods in a monopolistically competitive market. By following Christiano et al. (2008), the model includes expectations of future macroeconomic developments. As a matter of fact, shocks have two components: an unanticipated part and an anticipated element, meaning that, at time t, agents are provided with a signal about future macroeconomic developments at period t+n. In particular, model includes expectations on productivity, houses supply, cost of transforming output into capital, monetary rate, deviations from inflation goal and inflationary pressures. The first monetary regime case (i) is a typical Taylor rule targeting inflation and output. Secondly, there is a pure and strict anti-inflationary policy scenario (ii), in which central banker focuses only on inflation target and does not deviate from it. Thirdly, in order to smooth financial cycles, the central banker also targets financial variables case (iii) : either credit growth or fluctuations in housing prices. Finally, in order to increase credit in busts and decrease it in booms, a macro prudential countercyclical LTV ratio is introduced. It is modelled as dependent on GDP growth case (iv), credit variations scenario (v) and housing prices changes case (vi). 8

10 In order to rank the monetary regimes, authors use a welfare criterion based on conditional expectations of lenders and borrowers lifetime utility. Empirical structure: In calibrating the parameters, authors follow Iacoviello and Neri (2010). In opposition to similar studies, the authors do not want to distinguish among different shocks TFP vs financial shocks but simulate the economy response to fluctuations, whatever their source is. Empirical results: First, consider case (i) vs. case (ii). The results show that a typical Taylor rule yields much more stability and, under a merely rigid anti-inflationary monetary policy, both lenders and borrowers decrease their own welfare. In fact, case (ii) delivers greater instability because, in order to pursue strictly the inflation target, the policy rate is extremely volatile. Among other things, note that the simulations show that there is a trade off between the inflation and output variability, meaning that an interest rate rule that targets both output and inflation cannot decrease the inflation and output variability at the same time. Second case, scenario (iii) vs scenario (i). If the central banker considers also financial variables case (iii), then there are significant gains in terms of macroeconomic stability and both savers and debtors improve their welfare relative to a monetary policy based on a classical Taylor rule case (i). In particular, considering the credit growth as financial measure provides better results than considering housing prices: both agents welfare is greater and all macroeconomic variables, aside from inflation, show greater stability. The reason is that some inflation volatility helps flatting the real effects of variations in nominal debt and, as a result, leads to greater stability. Third, consider the macro prudential countercyclical LTV ratio. This rule is able to reduce output volatility; in particular, LTV ratio dependent on credit growth scenario (v) is more effective than LTV ratio dependent on either GDP growth case (iv) or housing prices changes scenario (vi). Anyway, the three scenarios show small differences in variability of housing prices, housing investments and inflation. Also, while LTV ratios responding on GDP growth and housing prices changes are not able to reduce the variability of loans and loan-gdp ratio, LTV tool responding on credit growth can do it. The bottom line is that the latter, given a shock, generates moderate responses in LTV ratio, while the response of LTV tool dependent on either GDP growth or housing prices is stronger, leading to more instability relative to LTV ratio based on credit growth. By using the welfare criterion, it is found that only LTV ratio dependent on credit growth improves both agents welfare, while the other two LTV tools increase savers welfare, but worsen borrowers one. The reason is that increasing volatility in credit flows, associated with LTV rules targeting either GDP growth or housing prices, drives up the fluctuations of debtors consumption and, as a consequence, decreases their welfare. Last but not least, using welfare analysis to rank monetary policy based also on financial variables scenario (iii) and a macro prudential policy targeting credit growth scenario (v) does 9

11 not provide consistent results: savers maximise their utility under case (iii), while borrowers are better off under scenario (v), wherein credit level is higher, allowing debtors to consume more nondurable goods and housing. Policy implications: Some policy conclusions emerge. First, a rigid anti-inflationary monetary policy mitigates business cycles fluctuations worst than what a typical Taylor rule does. Second, monetary regime taking into account credit growth drives up macroeconomic stability and improves welfare, meaning that monetary policy goals and business cycles stability do not show any trade off. Third, when macro prudential tools are introduced to fight boom and busts, countercyclical LTV measures reacting to credit growth are more effective in stabilising the credit cycle than a monetary policy responding to changes in credit aggregates. Anyway, since the former maximises borrowers welfare, while the latter improves lenders welfare, a consistent rank of the two regimes cannot be defined. Role of the Central Bank: In the model, the central bank is the monetary authority. It is assumed that the central banker can also implement macro prudential measures. As shown, the model suggests that the central banker should avoid a rigid anti-inflationary measure, while including credit growth considerations in implementing monetary policy helps stabilising the economy and turns out to be welfare improving for both lenders and borrowers. It also means that monetary policy goals and business cycles stability do not show any trade off. If macro prudential tools are introduced, countercyclical LTV measures reacting to credit growth are more effective in stabilising the credit cycle than what a monetary policy responding to changes in credit aggregates does. Anyway, since the former maximises borrowers welfare, while the latter improves lenders welfare, a consistent rank of the two regimes cannot be defined. Anyway, authors do not deal with the consequences of unifying the monetary and macro prudential authorities and do not face cooperation issue. Financial and Insurance markets: Financial sector is not segmented. Paper just models lenders and borrowers. II.IV MACRO PRUDENTIAL POLICY AND THE CONDUCT OF MONETARY POLICY Aim: Beau, Clerc and Mojon (2011) study the interactions between monetary and macro prudential policies and the need of coordination. In order to assess if and how the two policies have compounding, neutral or conflicting impacts on inflation, a DSGE model including financial frictions, heterogeneous agents and housing is used. The estimation is based on the US and Euro Area data from 1985 to The authors analyse four different policy regimes relying on: (i) monetary policy goals if the short term rate includes financial stability by leaning against credit bubbles ; (ii) the 10

12 existence of a macro prudential agency whose task is guaranteeing the financial stability by implementing supervision policies without impacting on the monetary policy rate. Four main results arise: macro prudential policy usually does not affect price stability; effects on inflation emerge only under financial shocks (shocks to asset prices and credit); under these shocks, an independent monetary policy focused on prices stability and an independent macro prudential authority leaning against the credit growth allow to reach the best outcome in terms of inflation stability; such an outcome is improved if monetary agency takes into consideration the macroeconomic effects deriving from macro prudential policy, meaning that coordination and sharing information on macro prudential policies are crucial to allow the central bank to keep the prices stable. Finally, given the theoretical framework above, the paper assesses the policy regimes in the US, UK and Euro Area. Theoretical framework: The framework is based on a DSGE model à la Iacoviello (2005) with residential investments, house prices and loans. The main elements of the model are the following: as in Liu et al. (2009), investment cycle is affected by housing prices; housing and preference shocks impact on agents utility functions; impatient households and entrepreneurs are borrowers whose credit constraint is binding in equilibrium, wherein borrowing is limited by the net present discount value of housing wealth; from this viewpoint, a positive financial shock is just a loosening of borrowers loan to value ratio (for example, due to an increase in the banking market competition or financial innovation); only impatient households are subject to the marginal utility of housing that, in turn, impacts on housing demand. The bottom line is that, contrary to Iacoviello (2005), the authors are focused on the interactions between a demand shock and a binding borrowing constraint within a framework of nominal debt indexation. What about government policies? As usual, the agency in charge of the monetary policy is the central bank, whose tool is the short term nominal interest rate. Central bank is subject to a typical loss function, in which it minimises the variance of inflation, output and interest rate itself. Macro prudential goal is leaning against the financial (credit growth) wind. It means that policymaker in charge of supervision is able to influence the loan to vale ratio of borrowers impatient households and entrepreneurs. Put it in other way, the macro prudential authority aims at flatting the deviations of credit growth from its steady state. Empirical structure: Data come from Euro Area and the US and cover the period Four policy regimes experiments are simulated and analysed: (1) Plain vanilla Taylor rule: the central bank sets the monetary policy tool the short term nominal interest rate by following a traditional Taylor rule. The only goal is the price stability and the interest rate goes up in responding both to inflation and output gap. No macro prudential policy is implemented. 11

13 (2) Lean against the financial wind Taylor rule: the central bank leans against the financial wind, meaning that the Taylor rule is augmented to allow interest rate to increase with credit expansion. (3) Independent macro prudential agency: the central bank and the supervision authority act separately by pursuing their own objectives and no coordination is implemented. (4) The central bank is modelled as in (2), but there is also an independent macro prudential authority that leans independently against the financial wind. Some different shocks are considered. Two of them are financial shocks credit supply and housing preferences shocks, while the others time preferences, productivity, investment, monetary policy and cost push shocks represent the traditional disturbances. Empirical results: By implementing a variance analysis, authors find that typical shocks cover 80% of the variance of the Euro Area inflation and more than 80% in the US. Also, under the five traditional shocks, inflation response is almost the same across four different policy regimes. These results confirm that in most fluctuations of business cycles, an independent supervision policy does not affect the price stability 3. Only under financial shocks, the impact on inflation differs across policy regimes. In both financial shocks, the inflation is more stable under regime (3), where an independent central bank pursuing only price stability goals is combined with an independent macro prudential supervision that leans against credit expansion: for instance, in the case of housing preference shocks, housing prices growth does not turn into a credit boom. Once authors have estimated and optimised the coefficients of the four policy rules relying on the US and Euro Area data, they analyse the outcomes in terms of variance of inflation, output, interest rate and credit across the policy regimes. The main outcome is that the policy (3) turns out to be the best regime to stabilise inflation, output and interest rate. Given the results above, cooperation seems to be useless: in order to assess the importance of coordination, the authors try to find the most efficient monetary policy constrained by a given macro prudential policy policy regime 1 vs policy regime 2. In other words, in terms of inflation and interest rate stabilisation, is it efficient for the central bank to take into account the macro prudential policy? The empirical results show that it could be more efficient to optimise the Taylor rule coefficients taking the macro prudential policy as given (case 1) rather than incorporating the macro prudential goals in the monetary policy (case 2), meaning that an independent supervision agency pursuing the financial stability does not affect necessarily the goal of price stability. Policy implications: In terms of central banking, some important policy implications arise. First, in most cases, inflation is not influenced by macro prudential policy; only under credit or asset price 3 Angelini, Neri and Panetta (2012) find similar outcomes. 12

14 shocks, supervision impacts on inflation dynamics and, in such cases, an independent macro prudential agency is very useful to stabilise the inflation. Second, the most efficient regime to keep inflation and output stable is a central bank following the Taylor rule, while adding a supervision agency improves such an outcome under financial shocks. Also, allowing the central bank to use the monetary tool to pursue the financial stability leads to suboptimal results. Third, in order to yield the best outcome in terms of inflation and output stabilisation, the central banker should be fully informed of the policy implemented by the supervision agency. Broadly speaking, the need of separate assignments arises, even if coordination is necessary. Finally, given the theoretical framework, the paper collects some policy advices for the FED, ECB and BoE. First, although the Financial Stability Oversight Council (FSOC) is the institution in charge of macro supervision in the US, the FED has a powerful role in pursuing macro prudential goals: it should be avoided to implement redundant supervision policies (scenario 4 risk). The authors argue that the institutional arrangement of the European Systemic Risk Board (ESRB) is separated by the ECB working, but the presence of single members central bankers both in ECB governing council and ESRB is likely to guarantee information sharing. In the UK, the new Financial Stability Committee works within the BoE, meaning that cooperation is likely to be guaranteed. At the same time, separation from the Monetary Committee should be able to avoid scenario 4 risk. Role of the Central Bank: The central bank is modelled in some different ways. Its goals can be traditional only monetary purposes by setting the interest rate to keep inflation and output stable (case 1). It can be in charge of both monetary and macro prudential goals (case 2), by leaning against the credit growth wind. It can focus only on monetary goals and acts independently from a supervision agency (case 3). It can be modelled as in case 2, but coexists with a macro prudential agency that also leans against credit growth wind (case 4). The most efficient regime to keep inflation and output stable is a central bank focusing on monetary goals, while adding a supervision agency improves such an outcome under financial shocks. Also, allowing the central bank to use the monetary tool to pursue the financial stability leads to suboptimal results, both with and without a supervision agency. Finally, in order to yield the best outcome in terms of inflation and output stabilisation, the central banker should be fully informed of the policy implemented by the supervision agency. Financial and Insurance markets: The lending sector is not differentiated. 13

15 II.V MONETARY AND MACRO PRUDENTIAL POLICIES Aim: Angelini, Neri and Panetta (2012) use a DSGE model with a stylized banking sector to verify the interaction between monetary and macro prudential policy. Under normal times the business cycle is supply-driven macro prudential supervision gains are quite modest. Also, the possible lack of cooperation between the monetary authority (the central bank) and the macro prudential agency could lead to suboptimal results. If financial shocks occur shocks to supply of loans macro prudential policy gives more benefits: the more the central bank will cooperate with the macro prudential agency and will broad its goals beyond the price stability, the more the macro prudential policy gains are greater in restoring the financial stability. Anyway, by implementing a welfare analysis of the different agents, two main results arise: no regime makes all agents better off and, from this viewpoint, a definitive rank is not possible; solid redistributive effects among agents occur both with supply and financial shocks. Theoretical framework: The theoretical template is based on a DSGE model by Gerali et al. (2010) with a stylized banking sector. The agents of the economy are entrepreneurs, (patient and impatient) households and banks. Patient households are savers who deposit their money in the banks. Impatient households and entrepreneurs borrow from the banks and face a binding credit constraint. Entrepreneurs produce consumer and investment goods using capital and labour supplied by households. Note that households are monopolistic in supplying the labour and nominal wages are assumed to be set by unions. Banks are monopolistically competitive and loans to impatient households and firms are banks assets, while liabilities are deposits and capital. Monopoly power allows the banks to set the interest rates on loans and deposits. Bank rates are sticky because some adjustment costs occur when they move loan and deposit rates. According to the balance sheet identity, loans are equal to deposits plus capital. Banks want to keep their capital-asset ratio close to an exogenous target v t representing a capital requirement imposed by the regulator the macro prudential tool of the model. From this point of view, two main elements are noteworthy: first, the total loans are the sum of the risk-weighted loans to entrepreneurs and households; second, banks capital can increase only through retained earnings. The equilibrium conditions show that if loans go up, the capital-asset ratio could turn out to be lower than the threshold v t, leading the banks to increase the lending rate. Of course, this drives down the credit demand and, in turn, decreases consumption and investments. In the model, monetary and macro prudential policy are independent and implemented by the central bank and a supervision authority, respectively: R t is the monetary policy rate (monetary policy tool) and affects both lending and deposit rates, while v t (macro prudential policy tool) impacts only on lending rates. It means that different effects on savers and borrowers could arise. The monetary policy rate follows a standard version of Taylor rule and the central bank uses a loss function to minimise the variance of inflation, output and interest rate itself. 14

16 The macro prudential tool v t is set by taking into account the output growth and choosing a countercyclical policy: capital-asset ratio increases in good times and goes down in a bust. The macro prudential authority uses a loss function to minimise the variance of loans to output ratio, output and v t itself. Empirical structure: The model calibration follows the prevalent literature for central banking, such as Ozlale (2003), Ilbas (2012), Ehrmann and Smets (2003). Anyway, robustness checks are implemented by setting alternative values for the two loss functions and for the macro prudential policy rule - v t -: results keep unchanged. The empirical analysis is based on three cases: (i) first, the macro prudential and monetary authorities are completely coordinated, meaning that the two policies are chosen jointly by a single policymaker with two tools (interest rate R t and capital requirement v t ), whose goal is minimising the variance of inflation, output, the loans to output ratio, and the changes in the tools themselves the two loss functions are merged. (ii) Second, no cooperation is assumed: the central bank minimises its own loss function and takes as given the macro prudential policy rule, while the macro prudential authority minimises its own loss function and takes as given the monetary policy rate. (iii) Third, macro prudential supervision is absent and there is only the monetary policy. Finally, both cases are studied by simulating technology normal times and financial shocks shocks to credit supply. Empirical results: First, a technology shock is considered. Under cooperation case, the macro prudential policy is countercyclical, because the more the output growth, the more the capital requirement is tightened. Under non cooperation scenario, the macro prudential policy turns out to be procyclical, while monetary policy is strongly countercyclical. The sum of loss functions is worst than scenario (i). In particular, under (ii), the variability of interest rate is 22 times greater and that of capital requirements is twice as great. It means that non cooperative case could lead to suboptimal results because of coordination problems monetary policy is countercyclical, while macro prudential supervision is not. Under non cooperation scenario, if a negative technology shock occurs, as the loans to output ratio goes up, the macro prudential agency tightens the capital requirements and worsens the recession. Note that loans to output ratio increases because bank credit decreases less than output, given that the impact on households and entrepreneurs borrowers is delayed. As a result, the central bank acts aggressively and, in turn, macro prudential authority keeps tightening. Under case (iii), results are better: volatility of interest rates and loans to output ratio declines strongly, meaning that under supply shocks macro prudential policy is not so useful. Second, a financial shock is taken into account. Under cooperation case, both monetary and macro prudential policy are strongly countercyclical. By analysing the loss functions, it is found that the central bank loses and macro prudential authority gains. The volatility of output, loans to output 15

17 ratio and capital requirement decreases, while the cost is a slight increase in the variability of inflation and monetary policy rate: the related interpretation is that central bank can deviate from its goal to support the financial stability. Under scenario (ii), macro prudential policy is still countercyclical, but the monetary policy reaction is weaker and the fluctuations of macro prudential variables are slightly greater than scenario (i). With respect to technology shock, there is no conflict because fluctuations in output and loan to output ratio are in the same direction. Under case (iii), the results worsen. The authors also implement a welfare analysis of each category of agents. Given the technology shock case, borrowers are better off under the cooperation regime. Savers are better off under non cooperative scheme. When a financial shock is considered, no regime makes all agents better off: savers reach the highest utility under cooperation, while borrowers do it under pure monetary policy regime. Policy implications: The model structure tries to capture some important policy aspects: first, the role of bank capital and credit supply in the financial crisis is represented; second, the capitalasset ratio is a good example of countercyclical capital buffer introduced by Basel 3; last but not least, a micro prudential regulator could be likely to require an undercapitalized bank to increase the capital-asset ratio, regardless of the adjustment is focused on reduction in asset or increase in capital. The former means that a micro prudential regulator could damage the economy: by following Hanson, Kashyap and Stein (2010) one can characterize the macro prudential approach to capital regulation as an effort to control the social costs associated with excessive balance-sheet shrinkage on the part of multiple financial institutions hit with a common shock, the model captures how the risk of assets shrinkage can be mitigated by the macro prudential policy. In normal times, the benefits of macro prudential policy are modest. In order to avoid suboptimal outcomes, cooperation between central bank and macro prudential agency should be guaranteed, even if monetary policy alone can do a better job in reaching both monetary and supervision purposes. The suboptimal result under (ii) depends on the fact that two independent authority have different goals: the central bank is focused on output and inflation stabilization, while the financial supervisor aims at keeping the credit supply stable. This is why conflicting policies can arise. If financial shocks are important factors in the economy, the benefits of implementing a macro prudential policy are substantial. The cooperation between monetary and macro prudential authority leads to a greater stability of macroeconomic variables: the cost is a very active monetary policy and bigger variability of inflation, meaning that central bank should deviate from its own objectives to support the financial stability. 16

18 The welfare analysis shows that no regime makes all agents better off. From a welfare viewpoint, it means that the optimal monetary and macro prudential policy depends on which kind of agent the policymaker wants to please. Role of the Central Bank: In the model, the central bank has only monetary objectives keep output and inflation stable and does not deal with supervision purposes. It is independent from the macro prudential authority. Nevertheless, cooperation between central bank and macro prudential agency is a key factor: in normal times, coordination allows to avoid suboptimal results, even if monetary policy alone can do a better job in reaching both monetary and supervision purposes. Under financial shocks, the central bank should deviate from its own objectives to support the financial stability. Financial and Insurance markets: There is only the banking sector, wherein banks are monopolistically competitive and set the lending and borrowing rates. II.VI THE MISSING POINTS Financial crisis in 2008 triggered the call for macro prudential policies. From this viewpoint, the literature shows some clear trends and three interesting results. Theoretically, DSGE models are implemented to analyse the interaction between macro prudential and monetary policy 4. Most of these models present a stylized banking sector to verify the interaction between monetary and macro prudential policy. The literature is coherent in admitting that under normal times the business cycle is supply-driven macro prudential supervision gains are quite modest (some simulations find losses in implementing macro prudential policies during traditional shocks), while, under financial disturbances, macro prudential policy gives more benefits. Why? As it has been correctly argued 5, the bottom line is that a financial shock tends to move the objectives of monetary and macro prudential policy (for instance, output and loan to output ratio, respectively) in the same direction, while the traditional shocks do the opposite. As a result, a key factor in implementing the optimal monetary and macro prudential measures is understanding the source of the shocks: in order to implement effective macro prudential policies, the policymaker should be able to distinguish between financial and real shocks. Consequently, the more the central bank as liquidity manager gains information advantages, the more its leading role in the definition and implementation of the macro supervision has to be established. 4 As shown, see N Diaye (2009), Kannan, Rabanal and Scott (2009), Lambertini, Mendicino and Punzi (2011), Beau, Clerc and Mojon (2011) and Angelini, Neri and Panetta (2012). 5 Angelini, Neri and Panetta (2012). 17